MDL
Nemotron 3 Ultra 550B A55B
NVIDIA
Verified against Artificial Analysis · Jul 13, 2026
Capability
87
percentile index
Price /M
$0.675 / $2.675
in / out
Context
1M
tokens
Avg score
60.2
7 benchmarks
AA Intelligence
37.8
index
Speed
189 tok/s
TTFT 0.9s
Released
Jun 4, 2026
About
Nemotron 3 Ultra 550B A55B is a open-weights model in the Nemotron family from NVIDIA. Benchmarked on 7 evals, averaging 60.2.
Benchmark scores · 7
avg 60.2 — higher is betterThe research behind this model
all papers mentioning it →Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning▲ 15 on HF · Jun 12, 2026ProCUA-SFT Technical Report▲ 9 on HF · Jun 15, 2026Diagnosing Overhead in Dispatch Operations: Cross-architecture ObservatoryMay 20, 2026Mixture of Experts for Low-Resource LLMsMay 17, 2026HEBATRON: A Hebrew-Specialized Open-Weight Mixture-of-Experts Language ModelMay 11, 2026
Mentions matched by name in title/abstract — from the arXiv + HF daily corpus.
Performance & price
Output speed189 tok/s
Latency (TTFT)0.9s
First answer13s
Price in / out$0.675 / $2.675 per 1M
Blended price$1.18 / 1M
Coding index49.3
Median across API providers, via Artificial Analysis. Blended = 3:1 input:output. For reasoning models, latency includes thinking time. Methodology →
History
full ledger →Capability over time
Append-only ledger — every observed change to this model's numbers.
API providers
Artificial Analysis$2.675/M